Searching for the Root of the Tree of Evil

“There are a thousand hacking at the branches of evil to one who is striking at the root”

Henry David Thoreau, Walden

The world is full of problems: Pain, poverty, illness, war, pollution, to pick a few among thousands. Many of us feel like we need to Do Something about these problems. There’s just one problem (sorry): There are so many problems that focusing on one specific problem feels wrong. Why choose that one? How dare you ignore all the other problems? Do those people suffering not matter to you? Can’t you hear their screams?

One response is the effective altruist’s shut up and multiply: loosely, find a framework that lets you rank problems against each other, and then work on the most important problems. Yes, war is hell, but solving it is not tractable or neglected. So you buy anti-malarial bednets and try your best to ignore the screams.

Another is to try to identify the root cause of all the problems, and work on that. Some popular candidates include metascience, “healing our collective trauma”, “awakening from the meaning crisis”, “solving intelligence and using it to solve everything else”.

Sometimes you can justify this on effective altruist grounds, but it’s also very neat psychologically: you don’t have to ignore the screams. Whenever a new problem comes to your attention, you can say “yes, my work is helping with this also”. You can channel that empathy into whatever you were already doing. And afterwards, you can sleep easy, knowing that you helped.

This is awfully convenient, and you might be tempted to dismiss it as cope, as a way to feel good about yourself without ever doing any real, object-level, devil-is-in-the-details kind of work. And yet… this meta-strategy has worked at least once before, and worked so well as to dwarf all the object-level charity ever wrought by human hands and hearts. I’m referring to Francis Bacon’s invention of science.

Quoting the historian and novelist Ada Palmer’s beautiful essay On Progress and Historical Change:

Then in the early seventeenth century, Francis Bacon invented progress.

If we work together — said he — if we observe the world around us, study, share our findings, collaborate, uncover as a human team the secret causes of things hidden in nature, we can base new inventions on our new knowledge which will, in small ways, little by little, make human life just a little easier, just a little better, warm us in winter, shield us in storm, make our crops fail a little less, give us some way to heal the child on his bed.  We can make every generation’s experience on this Earth a little better than our own.  […] Let us found a new method — the Scientific Method — and with it dedicate ourselves to the advancement of knowledge of the secret causes of things, and the expansion of the bounds of human empire to the achievement of all things possible.

There are many caveats: the Great Man theory of history is flawed, Bacon was just a spokesman for a broader community of British proto-scientists, science would have probably emerged anyway at some point, most progress comes from tinkerers and engineers rather than scientists proper. Nonetheless… has anyone ever been more right about anything than Francis Bacon was about the importance of the scientific method?

---

We now have an existence proof of a successful meta-intervention. Francis Bacon identified the root cause of most 17th century problems (our collective ignorance about the natural world) and worked to address them by developing and evangelizing the scientific method. What is the equivalent today?

Some say: still ignorance about the natural world! We need to double down on Science! If we just understood biology better we could cure all diseases, develop ever more nutritious foods, and find ways to arrest ecosystem collapse and environmental degradation. If we just understood physics better we could unlock new abundant sources of energy, solve poverty and climate change, and travel to the stars.

I don’t buy it. Science has already taught us how to be healthy: eat whole foods, exercise a lot, spend time outside, live in community. Yet we do less and less of all these things. Physics has already given us nuclear power, an incredibly cheap, reliable, safe, low-emission source of energy. We don’t use it because it’s politically unpopular. Our biologists and ecologists already know quite well how to protect ecosystems, as our national parks show. We just don’t care enough to do it.

Maybe the bottleneck is in distributing our scientific understanding more broadly? This is the position of those who see lack of education as the root of all problems. I don’t buy this one either. Our best analyses of education suggest that schools are ineffective at disseminating scientific knowledge. Even the knowledge that does get disseminated isn’t acted on: how many people at this point don’t know that regular exercise is the number one health intervention in the developed world?

All these examples suggest the bottleneck has something to do with people and their interactions. Perhaps we’re reached diminishing returns on natural science, but human and social sciences are still incredibly valuable. If we could only understand psychology and sociology and economics well enough, we could design interventions that convince people to act in accordance with their best interests, and that of society as a whole.

This feels compelling, but I can’t help but notice we’ve sunk hundreds of billions of dollars and millions of our brightest minds into this research over the last century, and yet… are we actually better at organizing our social life in accordance with human flourishing than the Victorians, the Qing, or the pre-conquest Lakota? In some ways yes, in some ways no. If we ignore the differences in material prosperity (largely due to our better science and engineering, not social science), I’d call it a wash at best. Admittedly, natural science took a long time to pay off: Bacon wrote Novum Organum in 1620 and the Royal Society was founded in 1660; the British economy started growing rapidly only around 1820. Perhaps all this psychology and social science research will pay off in the end. But I’m not holding my breath.

A final meta-intervention I’ll bring up to dismiss is the one exemplified by DeepMind’s founding mission statement: “solve intelligence. use it to solve everything else”. This has been the ideology of Silicon Valley for the last year or so, and seems poised to become the ideology of financial capitalism and the military-industrial complex as a whole. Instead of solving object-level problems, we are funneling all our surplus capital, our best analytical minds, and (soon) all our surplus energy towards increasing the supply of raw silicon-based intelligence in the hopes that it will solve our problems for us. I’ll address this more fully in a future essay, but briefly I have the same reaction to it as Fouché had to the murder of the Duc D’Enghien: “It is worse than a crime; it is a mistake”. Intelligence is merely an accelerant; it will amplify both the best and the worst trends in society; more intelligence helps only if you believe we are on a good trajectory on net, and our main problem is that we’re not getting there faster enough.

Enough criticism - what do I think the real root of all evil is? As you might have guessed from the above, I believe it’s our inability to understand and cooperate with each other at scale. There are different words for the thing we need more of: trust. social fabric. asabiyyah. attunement. love. But all these words are deceptively cozy, suggesting we just need to Retvrn to our tribal and relational instincts and all will be okay. This is a dangerous illusion. Our old social technologies for cooperation did not scale to the complexity and enormity of the modern world, and were replaced by global capital markets and massive state and corporate bureaucracies. We need to find the principles underlying these cozy words, and find a way to make them scale. Much like Bacon and Boyle built a Knowledge Machine that takes in irrational argumentative academics and produces scientific knowledge; much like Hinton and Schmidhuber and Sutskever built a Learning Machine that takes in oceans of data and computation and produces raw intelligence; we need to build a Cooperation Machine that takes in atomized people and raw intelligence and produces mutual understanding and harmonious collective action.

Thanks to Richard Ngo for his inimitable combination of encouragement and skepticism that helped sharpen the ideas in this essay.